Wake by Robert J. Sawyer is the first novel of a trilogy, it came out in 2009, Watch, the second book, came out this year, and Wonder will come out in 2011. Sawyer calls them the WWW Trilogy, and it has a rather slick web site, with the best production values I’ve ever seen promoting a SF novel. Personally, I found Wake as exciting as when I first discovered science fiction back in the 1960s, when I was a kid. And it’s up for the Hugo this year, so I figure Penguin knows it has a great story and its hitting warp ten to promote it.
Wake is not marketed as a YA novel, but it could have been. The main character is Caitlin Decter, a fifteen year old blind girl, who is a math wiz, computer geek, engaging blog writer, and extremely precocious. This reminds me tremendously of the Heinlein juveniles from the 1950s, and in particular Holly from “The Menace From Earth.” Like the Heinlein juveniles, Wake is chock full of educational tidbits. And Wake is the kind of novel you don’t want to put down.
Classic SF Theme: Intelligent Computers
It’s getting harder and harder for science fiction writers to come up with completely new science fictional ideas, so what we often see is a writer taking on a classic theme and having a go at evolving past ideas. Wake follows in the tradition of many fictional computers, but in particular ones about a computer becoming conscious in front of one person. These are the just the ones I’ve read, there are many others.
- 1966 – The Moon is a Harsh Mistress by Robert A. Heinlein
- 1972 – When H.A.R.L.I.E. Was One by David Gerrold
- 1995 – Galatea 2.2 – Richard Powers
- 2009 – Wake by Robert J. Sawyer
Sawyer goes further then earlier writers in trying to imagine how an artificial mind would evolve and what it would perceive as it came into being. Sawyer weaves blindness and Helen Keller, autism, apes that do sign language, Julian Jaynes’ the bicameral mind, and other explorers of consciousness into the story in a very effective way.
One reason why I love this novel so much is because I’ve been writing a novel in my head about this subject for years. It’s not likely I’ll ever become a real novelist, but if I do, I’ll have to take the concept further than Sawyer, and that’s a good challenge.
Go read Wake. End of review.
Now I want to discuss what Sawyer is really writing about. Sawyer supposes that the Internet could evolve into a self-aware mind. That idea isn’t new, but what Sawyer does with Wake is make his case for it with series of suppositions that are wrapped in a page turning novel. In other words, he has a bunch of wild theories that he gets readers to think about one at a time.
What I’d like to do is discuss these ideas but hopefully without hurting anyone’s enjoyment of the story, but I recommend you not read beyond this point if you haven’t read Wake yet and want to get the full impact of its excitement.
Sawyer’s first theory is the emerging web mind will go through a stage much like what Helen Keller went through before she discovered language. Sawyer indirectly explores this stage in a number of ways, including quotes and references to Helen Keller, a subplot about signing apes, and references to the book The Origins of Consciousness in the Breakdown of the Bicameral Mind by Julian Jaynes, a book I found very exciting when it came out back in 1976.
But I think Sawyer is missing a piece of the puzzle, one I got from On Intelligence by Jeff Hawkins. Hawkins thinks our consciousness emerged out of a pattern recognition processor that we call the neocortex. Sawyer uses cellular automata as his theoretical model, but I’m not sure that will work. Cellular automata create patterns, but do they recognize them? I’m not sure the Internet can generate a consciousness in its current design. Oh, the Internet will make a fine nervous system for such a web being, or beings, but I think another type of device will need to be built first, and that’s a multilayer pattern recognizer that’s as good or better than our neocortex.
So far, all the writers exploring this theme have assume that when computers reach a critical mass a consciousness will spontaneously arise out of the complexity. I doubt that completely. I think at least three components are needed for self-aware consciousness: pattern recognition, mind and language. I don’t think any of these exist in the internet, or supercomputers. I think mind evolves out of pattern recognition, and self-awareness evolves out of mind, with the development of language.
Atoms and molecules have early stages of pattern recognition, but as life arises out of non-life, sense organs develop that seek out patterns in reality. Most organisms are so highly adapted to specific patterns that they will die off if they can’t find them. Evolutionary adaptation is the ability of organisms to explore and take advantage of new patterns. I believe the mind grows out of this process, and there are different kinds of minds. A dog, cat, dolphin and chimp all have minds. We aren’t sure how much they perceive, or if they have much self-awareness, but they do have minds. Language studies in dolphins and chimps hint that maybe these animals are self-aware and have identities, maybe far more than our egos want to believe, but I think their consciousnesses are limited by the state of their language abilities. I think signing will add consciousness to apes.
For an AI computer to develop a mind, I think it needs to have a focus on reality that is processed through a pattern recognition device, and then a language needs to be linked to the patterns. At first, I thought Sawyer was going to have the web mind see out of Caitlin’s artificial eye, so as the device taught her mind to see, the web mind would also learn to see, and with another fictional piece of technology, learn a language. Instead Sawyer imagines an inner world where the web is visible. I don’t buy that at all. It’s leftover fluff from cyberpunk novels. Why invent a new reality to observe, when the internet mind has millions of eyes on our reality?
Now this brings up some interesting questions about AI minds. If a web mind has millions of web cameras at its disposal, will it think think like it has millions of eyes? Or will it’s consciousness move from camera to camera and peer out at single points of reality? Omniscient life would be tough, don’t you think? I tend to believe, and I only have limited knowledge to think otherwise, that an AI mind will emerge from a limited environment. Some scientist will raise up an AI mind by teaching it to see and hear while learning a language.
But what will a hive AI mind be like? Let’s say anyone in the future can go down to Radio Shack and buy an artificial neocortex to add to their computer system and bring up an AI child. If all of these AI minds are connected by the Internet it will be like a race of telepathic beings. Now, wouldn’t that be a far out science fiction story? I still haven’t read Watch, so who knows what will happen.
JWH – 5/13/10